Continual selection of scenarios based on identified tags describing contextual environment of a user for execution by an artificial intelligence model of the user by an autonomous personal companion

Patent metadata

Field	Value
Publication number	US-2019065960-A1
Application number	US-201715684830-A
Country	US
Kind code	A1
Filing date	Aug 23, 2017
Priority date	Aug 23, 2017
Publication date	Feb 28, 2019
Grant date	—

Abstract

An autonomous personal companion executing a method including capturing data related to user behavior. Patterns of user behavior are identified in the data and classified using predefined patterns associated with corresponding predefined tags to generate a collected set of one or more tags. The collected set is compared to sets of predefined tags of a plurality of scenarios, each to one or more predefined patterns of user behavior and a corresponding set of predefined tags. A weight is assigned to each of the sets of predefined tags, wherein each weight defines a corresponding match quality between the collected set of tags and a corresponding set of predefined tags. The sets of predefined tags are sorted by weight in descending order. A matched scenario is selected for the collected set of tags that is associated with a matched set of predefined tags having a corresponding weight having the highest match quality.

First claim

What is claimed is:

1 . A method comprising:
capturing data related to behavior of a user using an autonomous personal companion providing services to the user; analyzing the data to identify one or more patterns of user behavior in the data from a plurality of predefined patterns, wherein each of the plurality of predefined patterns is associated with a corresponding predefined tag, wherein the plurality of predefined patterns is generated from a deep learning engine; classifying the identified patterns as a collected set of tags, wherein tags in the collected set are associated with the one or more identified patterns; comparing the collected set of tags to each of a plurality of sets of predefined tags associated with a plurality of scenarios, wherein each scenario corresponds to one or more predefined patterns of user behavior and a corresponding set of predefined tags; assigning a weight to each of the sets of predefined tags based on the comparing, wherein each weight defines a corresponding match quality between the collected set of tags and a corresponding set of predefined tags; sorting the plurality of sets of predefined tags by corresponding weights in descending order; and selecting a matched scenario to the collected set of tags, wherein the matched scenario is associated with a matched set of predefined tags having a corresponding weight with the highest match quality.

2 . The method of claim 1 , further comprising:
providing the captured data as input into a matched algorithm of the matched scenario that is executed to determine a behavior associated with the personal companion; and performing one or more actions based on the determined behavior, wherein at least one action includes moving the personal companion.

3 . The method of claim 1 , further comprising:
accessing data related to monitored behavior of the user; accessing data related to monitored behavior of a plurality of users; determining the plurality of predefined patterns predicting behavior of the user based on the collected data.

4 . The method of claim 1 , further comprising:
collecting the data on a continual basis; determining a change of context based on the collected tags that are updated; comparing the updated collected set of tags to each of the plurality of sets of predefined tags associated with a plurality of scenarios; assigning an updated weight to each of the sets of predefined tags based on the comparing; sorting the plurality of sets of predefined tags by the corresponding updated weights in descending order; and selecting an updated matched scenario to the updated collected set of tags that is associated with an updated matched set of predefined tags having a corresponding updated weight with the highest match quality.

5 . The method of claim 1 , further comprising:
setting an expiration period for each of the plurality of algorithms of the plurality of scenarios.

6 . The method of claim 1 , further comprising:
determining audio data from the captured data based on at least one of the collected tags; classifying the audio data into one of command speech, background scenario speech, and conversation speech; and aligning the result with the classified audio data.

7 . The method of claim 1 , wherein the execution of the matched algorithm further comprises:
determining an emotional state of the user based on at least one of the collected tags; and providing a therapy based on the emotional state as one of the actions.

8 . The method of claim 1 , wherein the execution of the matched algorithm further comprises:
determining an emotional state of the user based on at least one of the collected tags; and providing animation of an object reflecting the emotional state as one of the actions.

9 . The method of claim 2 , further comprising:
determining when moving that the personal companion is approaching a private zone in physical space; and preventing the personal companion from entering the private zone.

10 . The method of claim 2 , further comprising:
positioning the personal companion closer to the user when performing the moving.

11 . The method of claim 2 , further comprising:
following the user when performing the moving.

12 . The method of claim 2 , further comprising:
positioning the personal companion when moving to better project images from the personal companion onto a displayable surface; and projecting the images as one of the actions.

13 . The method of claim 2 , wherein the matched algorithm selects the one or more actions to be performed from a plurality of possible actions.

14 . The method of claim 2 , further comprising:
starting a gaming application for play by the user as one of the actions.

15 . A non-transitory computer-readable medium storing a computer program for implementing a method, the computer-readable medium comprising:
program instructions for analyzing the data to identify one or more patterns of user behavior in the data from a plurality of predefined patterns, wherein each of the plurality of predefined patterns is associated with a corresponding predefined tag, wherein the plurality of predefined patterns is generated from a deep learning engine; program instructions for classifying the identified patterns as a collected set of tags, wherein tags in the collected set are associated with the one or more identified patterns; program instructions for comparing the collected set of tags to each of a plurality of sets of predefined tags associated with a plurality of scenarios, wherein each scenario corresponds to one or more predefined patterns of behavior and a corresponding set of predefined tags; program instructions for assigning a weight to each of the sets of predefined tags based on the comparing, wherein each weight defines a corresponding match quality between the collected set of tags and a corresponding set of predefined tags; program instructions for sorting the plurality of sets of predefined tags by corresponding weights in descending order; and program instructions for selecting a matched scenario to the collected set of tags, wherein the matched scenario is associated with a matched set of predefined tags having a corresponding weight with the highest match quality.

16 . The computer-readable medium of claim 15 , further comprising:
program instructions for providing the captured data as input into a matched algorithm of the matched scenario that is executed to determine a behavior associated with the personal companion; and performing one or more actions based on the determined behavior, wherein at least one action includes moving the personal companion.

17 . The computer-readable medium of claim 15 , further comprising:
program instructions for collecting the data on a continual basis; program instructions for determining a change of context based on the collected tags that are updated; program instructions for comparing the updated collected set of tags to each of the plurality of sets of predefined tags associated with a plurality of scenarios; program instructions for assigning an updated weight to each of the sets of predefined tags based on the comparing; program instructions for sorting the plurality of sets of predefined tags by the corresponding updated weights in descending order; and program instructions for selecting an updated matched scenario to the updated collected set of predefined tags that is associated with an updated matched set of tags having a corresponding updated weight having the highest match quality.

Assignees

• Sony Interactive Entertainment Inc

Inventors

• Taylor Michael
• Fernandez-Rico Javier
• Bashkirov Sergey
• Yoo Jaekwon
• Chen Ruxin

Classifications

• G06N20/00

  Machine learning · CPC title

• G06N3/086 Primary

  using evolutionary algorithms, e.g. genetic algorithms or genetic programming · CPC title

• G06N3/008 Primary

  based on physical entities controlled by simulated intelligence so as to replicate intelligent life forms, e.g. based on robots replicating pets or humans in their appearance or behaviour · CPC title